Carpentries: R Intro PARTII Data Visualization
Alma Parada
10/28/2022
Load packages we will be using today
We could load only when we need it, but I like knowing what packages we need, so that I can install if I don’t have them.
library(tidyverse)
library(patchwork)
library(readxl)
library(ggplot2) #We will primarily work with ggplot2 today, but lots of optionsRead in data and prepare
#Read in the data
sea_chem<-read.table("sim_data.csv",header=T,sep=",",na.strings="NA",row.names=1)
samp_data<-read_xlsx("sim_data.xlsx","samples")
sea_comb<-merge(sea_chem,samp_data,by.x="row.names",by.y="sample_id")
#View(sea_comb) #note that the row names are now a column, we could change this, but for dataframes it's usually better to keep this valuable information in a column
#Create new column to group by based on the values of another
sea_comb$NH4_stat <- ifelse(sea_comb$NH4_nanoM<20,"depleted","replete")
unique(sea_comb$NH4_stat)## [1] "replete" "depleted"# Check the data type and some basic stats
str(sea_comb)## 'data.frame': 100 obs. of 8 variables:
## $ Row.names: 'AsIs' chr "S001" "S002" "S003" "S004" ...
## $ NH4_nanoM: num 379 312 405 298 316 ...
## $ NO3_nanoM: num 517 435 459 336 401 ...
## $ Bact : int 3210498 3607997 3393192 3316379 3759052 3255187 3976605 3749163 3538062 3942572 ...
## $ Virus : int 35013025 36342384 37485073 35002750 39974307 33830160 32538349 38923467 38681229 37545017 ...
## $ location : chr "coastal" "coastal" "coastal" "coastal" ...
## $ depth : num 10 10 10 10 10 10 10 10 10 10 ...
## $ NH4_stat : chr "replete" "replete" "replete" "replete" ...summary(sea_comb)## Row.names NH4_nanoM NO3_nanoM Bact
## Length:100 Min. : 0.5491 Min. : 327.3 Min. : 280554
## Class :AsIs 1st Qu.: 13.1097 1st Qu.: 441.5 1st Qu.:1391627
## Mode :character Median : 52.6819 Median : 529.2 Median :2212338
## Mean :113.3298 Mean :1086.9 Mean :2122338
## 3rd Qu.:134.5551 3rd Qu.:1060.8 3rd Qu.:2908654
## Max. :449.6311 Max. :3531.2 Max. :3976605
## Virus location depth NH4_stat
## Min. : 2820997 Length:100 Min. :10.0 Length:100
## 1st Qu.:13928311 Class :character 1st Qu.:10.0 Class :character
## Median :21949386 Mode :character Median :50.0 Mode :character
## Mean :20726538 Mean :30.8
## 3rd Qu.:26669133 3rd Qu.:50.0
## Max. :39974307 Max. :50.0#Could also use the dplyr package (part of tidyverse) to summarize
sea_comb %>% #start with the dataset "sea_comb", then
group_by(NH4_stat) %>% # group by the variable NH4_stat, then
count() %>% # count the number of observations, then
ungroup() #remove the grouping## # A tibble: 2 × 2
## NH4_stat n
## <chr> <int>
## 1 depleted 35
## 2 replete 65sea_comb %>%
group_by(location) %>%
summarise(mean_NH4 = mean(NH4_nanoM),
sd_NH4 = sd(NH4_nanoM),
mean_NO3 = mean(NO3_nanoM),
sd_NO3 = sd(NO3_nanoM))## # A tibble: 4 × 5
## location mean_NH4 sd_NH4 mean_NO3 sd_NO3
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 coastal 351. 45.8 446. 56.2
## 2 gyre 3.70 1.79 603. 47.0
## 3 offshore 72.4 6.40 443. 41.5
## 4 openOcean 26.3 9.15 2856. 359.# you could then convert this more easily into a table and export if desiredExercise 1
#Create a summary table of the bacterial and viral counts, include the mean and standard error for each location and depth
sea_comb %>%
group_by(location, depth) %>%
summarise(mean_Bact = mean(Bact),
se_Bact = sd(Bact)/sqrt(n()),
mean_Virus = mean(Virus),
se_Virus = sd(Virus)/sqrt(n()))## `summarise()` has grouped output by 'location'. You can override using the
## `.groups` argument.## # A tibble: 8 × 6
## # Groups: location [4]
## location depth mean_Bact se_Bact mean_Virus se_Virus
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 coastal 10 3559338. 89821. 36047571. 743702.
## 2 coastal 50 2737640. 127603. 26075276. 1266930.
## 3 gyre 10 794588 50143. 7747159. 468649.
## 4 gyre 50 691567. 62726. 6393958. 683433.
## 5 offshore 10 2820644. 148187. 25350993. 1130446.
## 6 offshore 50 2608197. 100502. 25468211. 904790.
## 7 openOcean 10 1936196. 91479. 17877316. 728636.
## 8 openOcean 50 1878338. 74250. 21168506. 918826.Bar Graphs, Histograms, Scatterplots
#Create a simple bar graph
ggplot(sea_comb, mapping = aes(x = NH4_stat)) +
geom_bar()
#many options for customization will be available if you either call the help pages or hit tab
#Modify bar chart to show percent of samples that are in each group
bar<- ggplot(data = sea_comb, mapping = aes(x = NH4_stat)) +
geom_bar(aes(y = (..count..)/sum(..count..))) +
scale_y_continuous(name = "Percent", labels=scales::percent)
bar
#Create a simple histogram
ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#Change the binwidth
hist_NH4<- ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram(binwidth = 10) # each bin covers 10nM
#Create a density plot
ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_density()
#Histogram & density
# Histogram with kernel density
ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram(aes(y = ..density..)) +
geom_density()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#Create a scatterplot
scat_Micro <- ggplot(sea_comb,aes(x=Bact,y=Virus)) +
geom_point()
scat_Micro
Plotting multiple figures together
#using bultin command par() & layout() cannot be used with ggplot2
attach(sea_comb)
# figures arranged in 3 rows and 1 columns
par(mfrow=c(3,1))
hist(NH4_nanoM)
hist(NO3_nanoM)
plot(Bact,Virus)
#Can change the margins of the figure
#mar=c(b,l,t,r) or mai for inches
par(mfrow=c(3,1),mar=c(2,2,2,0))
hist(NH4_nanoM)
hist(NO3_nanoM)
plot(Bact,Virus)
dev.off()## null device
## 1par(mfrow=c(3,1))
hist(NH4_nanoM)
hist(NO3_nanoM)
plot(Bact,Virus)
par("mar")## [1] 5.1 4.1 4.1 2.1dev.off()## null device
## 1#1 figure in row 1 and 2 figures in 2
layout(matrix(c(1,1,2,3),nrow=2,ncol=2,byrow=T))
hist(NH4_nanoM)
hist(NO3_nanoM)
plot(Bact,Virus)
#2 figure in row 1 and 1 figures in 2
layout(matrix(c(1,2,3,3),nrow=2,ncol=2,byrow=T),
heights=c(1,2),
widths=c(2,2)) # figs 1 & 2 in row 1, fig 3 in row 2
hist(NH4_nanoM)
hist(NO3_nanoM)
plot(Bact,Virus)
detach(sea_comb)
#Use patchwork package to plot graphs in one figure
(bar + hist_NH4) / scat_Micro
Saving figures
Color and Label Customizations
#Customize our histogram
hist_NH4 <- ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram(binwidth = 10, fill="red") # each bin covers 10nM
#Use built-in themes to further modify
hist_NH4 + theme_classic() #Change this up and see how it affects the graph
#Customize labels
#there are different ways to change the labels and breaks
hist_NH4 + scale_x_continuous(name = "Ammonium Concentration (nM)") +
scale_y_continuous(name = "Number of Samples",
breaks = c(0,5,10,15,20,25)) +
theme_light()
#Common mistake is to use the wrong scale
#h + scale_x_continuous(name = "Ammonium Concentration (nM)") +
# scale_y_discrete(name = "Number of Samples",
# breaks = c(0,5,10,15,20,25)) +
#theme_light()
#Customize a more complex figure
comb_NH4<-ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram(aes(y = ..density..),
binwidth = 10) +
geom_density()+
#Next let's change the amount of space data will take up
scale_y_continuous(name="Density",expand=expansion(mult=c(0.01,.01))) + #come back and adjust to low end from 0 to 0.1 , 0.01
scale_x_continuous(name="Ammonium (NH4, nM)",expand=expansion(mult=c(0.01,0.01))) #comback and adjust the sides to 0.01, 0.01
comb_NH4
#Take more control over the theme of the plot
comb_NH4 +
theme(panel.background = element_rect("white", "black", 2, "solid", "black"),
panel.grid.major = element_line(size = 0.2, linetype = 'dashed',
colour = "gray"))
#Let's add some color!
#can use hexadecimal color codes instead of names http://www.cookbook-r.com/Graphs/Colors_(ggplot2)/
comb_NH4_v2 <-
ggplot(sea_comb, aes(x = NH4_nanoM)) +
geom_histogram(aes(y = ..density..),
binwidth = 10,
colour="#999999", fill="#FFFFFF") + #changes the color of the line and the fill of the bars
geom_density(lwd = 1,
linetype = 2,
colour = 2) + #Changes the density line width, line type, and color
scale_y_continuous(name="Density",expand=expansion(mult=c(0.01,.01))) +
scale_x_continuous(name="Ammonium (NH4, nM)",expand=expansion(mult=c(0.01,0.01))) +
theme_classic(base_size = 16) #modifies a pre-existing theme
comb_NH4_v2
#can use hexadecimal color codes instead of names http://www.cookbook-r.com/Graphs/Colors_(ggplot2)/Exercise 2
#1 change the bar colors of the bar chart to a blue of your choice using it's hex code
#2 change the theme of the newly colored bar chart
#3 Change the x & y labels of the chart in #2
#4 plot the new bar chart and the original in one figure and save it
new<- ggplot(sea_comb, mapping = aes(x = NH4_stat)) +
geom_bar(fill="#6699CC")+
theme_bw() +
xlab("Nitrogen Status") +
scale_y_continuous(name="# of Samples")
comp_bars<- bar + new
ggsave("comp_bar.png",comp_bars,dpi=100)## Saving 7 x 5 in imageCustomize graphs by grouping variables: color and point shapes
#Need to prepare grouping variables, typically need to tell ggplot2 a variable is a factor
unique(sea_comb$location) #note output## [1] "coastal" "offshore" "openOcean" "gyre"sea_comb$location<-as.factor(sea_comb$location)
unique(sea_comb$location) #note order of levels## [1] coastal offshore openOcean gyre
## Levels: coastal gyre offshore openOceanscat_Nit2<-ggplot(sea_comb,aes(x=NO3_nanoM,y=NH4_nanoM,col=location)) + #color by location
geom_point()
scat_Nit2 #uses default rainbow colors
#order of location in legend could better represent our collection strategy
#Indicate the order the factors should be in
sea_comb$location<-factor(sea_comb$location,levels=c("coastal","offshore","openOcean","gyre"))
scat_Nit3<-ggplot(sea_comb,aes(x=NO3_nanoM,y=NH4_nanoM,col=location)) +
geom_point()+
scale_color_manual(values = c("#ca7dcc",
"#1b98e0",
"#353436",
"darkred")) #indicate the colors each level should use
scat_Nit3
#Change point shape to identify a second grouping variable (depth)
scat_Nit4<-ggplot(sea_comb,aes(x=NO3_nanoM,y=NH4_nanoM,col=location)) +
geom_point(aes(shape=as.factor(depth),color=location),size=2)+ #note I changed the depth to a factor in place
scale_shape_manual(values=c(3,16)) +
scale_color_manual(values = c("#ca7dcc",
"#1b98e0",
"#353436",
"darkred")) +
theme_bw()
#the legend labels cold be better!
scat_Nit4 + labs(color = "Location",shape="Water depth (m)") # Show them a new thing, change legend labels
#we could also move the legend
scat_Nit4 + labs(color = "Location",shape="Water depth (m)") +
theme(legend.position = "top") #values allowed, left, top, right, bottom
scat_Nit4 + labs(color = "Location",shape="Water depth (m)") +
theme(legend.position = c(0.7,0.7)) 
#give coordinates, values 0-1,
#c(0,0) corresponds to the “bottom left”
#c(1,1) corresponds to the “top right” position.
scat_Nit4 + labs(color = "Location",shape="Water depth (m)") +
theme(legend.position = c(0.7,0.7), legend.box="horizontal") #horizontally splits the legends
scat_Nit4 + labs(color = "Location",shape="Water depth (m)") +
theme(legend.position = c(0.7,0.7), legend.box="horizontal") +
guides(color=guide_legend(order=2),
shape=guide_legend(order=1)) #change the order of each guide
Exercise 3
#Modify the dataset and code to produce a NO3 x NH4 scatterplot that does the following
#1 Use color to identify NH4 replete samples
#2 Use shape to identify NO3 replete samples where NO3 > 2microMolar
#3 Improve the labels for the graph
#4 Choose a different theme for the graph
#First need to create a new grouping variable
sea_comb$NO3_stat<-ifelse(sea_comb$NO3_nanoM>2000,"replete","depleted")
#Create a ggplot scatter plot, x =NO3, y=NH4, color by NH4_stat, but the shape is the new NO3_stat variable
p<-ggplot(sea_comb,aes(x=NO3_nanoM,y=NH4_nanoM,col=NH4_stat)) +
geom_point(aes(shape=NO3_stat,color=NH4_stat),size=2)+
scale_shape_manual(values=c(17,16)) +
scale_color_manual(values = c("red",
"blue"))
p + xlab("Nitrate (nM)") +
ylab("Ammonium (nM)") +
theme_light(base_rect_size = 2) +
labs(color = "Ammonium",shape="Nitrate") +
theme( legend.position = c(0.7,0.7), legend.box="horizontal",
legend.background = element_rect(size=1,linetype="solid",color="black"))
Detecting patterns in scatterplots
ggplot(sea_comb, aes(x = Bact, y = Virus)) +
geom_point() +
geom_smooth(method = "lm") #draws a linear regression line, by default includes confidence envelope## `geom_smooth()` using formula 'y ~ x'
#Indicate groups within the scatterplot
ggplot(sea_comb, aes(x = Bact, y = Virus, color=as.factor(depth))) +
geom_point() +
geom_smooth(method = "lm") +
scale_colour_discrete(name = "Water Depth (m)",
labels = c("10", "50"))## `geom_smooth()` using formula 'y ~ x'
#Customize the colors using built in pallettes
ggplot(sea_comb, aes(x = Bact, y = Virus,color=as.factor(depth))) +
geom_point() +
geom_smooth(method = "lm") +
scale_color_brewer(palette = "Set1",
name = "Water Depth (m)",
labels = c("10", "50")) # http://www.cookbook-r.com/Graphs/Colors_(ggplot2)/## `geom_smooth()` using formula 'y ~ x'
Exercise 4
#create a scatterplot of Bacteria and Virus counts, and
#1 group the samples by location
#2 use the scale_* functions to rename the axes and customize the group colors
#3 fit a line to each group and remove the confidence envelope
unique(sea_comb$location)## [1] coastal offshore openOcean gyre
## Levels: coastal offshore openOcean gyreggplot(sea_comb, aes(x = Bact, y = Virus,color=location)) +
geom_point() +
geom_smooth(method=lm,se=F) +
scale_color_brewer(palette = "Dark2",
name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean", "Gyre")) ## `geom_smooth()` using formula 'y ~ x'
#4 what could we change to color by location but fit a line to all data points or by depth
ggplot(sea_comb, aes(x = Bact, y = Virus)) +
geom_point(aes(color=location)) +
geom_smooth(method=lm,se=F) +
scale_color_brewer(palette = "Dark2",
name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean", "Gyre")) ## `geom_smooth()` using formula 'y ~ x'
ggplot(sea_comb, aes(x = Bact, y = Virus,linetype=as.factor(depth))) +
geom_point(aes(color=location)) +
geom_smooth(method = 'lm',color="black")## `geom_smooth()` using formula 'y ~ x'
#Extra
#to get the formula and r2 need to do some work, https://stackoverflow.com/questions/7549694/add-regression-line-equation-and-r2-on-graph
BvV<-lm(Bact~Virus,sea_comb)
summary(BvV)##
## Call:
## lm(formula = Bact ~ Virus, data = sea_comb)
##
## Residuals:
## Min 1Q Median 3Q Max
## -898531 -298354 -62101 199847 1307823
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.600e+05 1.122e+05 2.317 0.0226 *
## Virus 8.985e-02 4.901e-03 18.335 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 476500 on 98 degrees of freedom
## Multiple R-squared: 0.7743, Adjusted R-squared: 0.772
## F-statistic: 336.2 on 1 and 98 DF, p-value: < 2.2e-16lm_eqn <- function(df){
m <- lm(Bact ~ Virus, df);
eq <- substitute(italic(y) == a + b %.% italic(x)*","~~italic(r)^2~"="~r2,
list(a = format(unname(coef(m)[1]), digits = 2),
b = format(unname(coef(m)[2]), digits = 2),
r2 = format(summary(m)$r.squared, digits = 3)))
as.character(as.expression(eq));
}
p<-ggplot(sea_comb, aes(x = Bact, y = Virus)) +
geom_point(aes(color=location)) +
geom_smooth(method=lm,se=F) +
scale_color_brewer(palette = "Dark2",
name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean", "Gyre"))
p + geom_text(x=1e6,y=3.5e7,label=lm_eqn(sea_comb),parse=T)## `geom_smooth()` using formula 'y ~ x'
Transforming data
Our data is currently in wide format
- row = participant or sample, column = variable
- gplot and computers don't really like this human readable formatWe can transform (reshape) our data to long-format
- row = observation
- each sample will have multiple rows, one for each observation
- each row is unique when looking at the combination of each measurement
- each variable that shares an axis will typically be in the same column#show how to "hide" R markdown results
# many ways to reshape data from wide <-> long
head(sea_comb)
#reshape {stats}
reshape(sea_comb,idvar="Row.names",
varying=c("NH4_nanoM","NO3_nanoM"),v.name=c("Conc_nM"),
times=c("NH4_nanoM","NO3_nanoM"),
direction="long")
#tidyverse, pivot_longer
sea_long<-pivot_longer(sea_comb,
cols = c("NH4_nanoM","NO3_nanoM"),
names_to="Nitrogen",
values_to="Conc_nM")
head(sea_long)
#modify column values
sea_long$Nitrogen<-vapply(strsplit(sea_long$Nitrogen,"_"), `[`, 1, FUN.VALUE=character(1))
head(sea_long)Visualizing our reshaped data
#Splitting figures using grouping variables with facet_grid
ggplot(sea_long, aes(y = Conc_nM,x=location)) +
geom_point() +
facet_grid(~Nitrogen)
ggplot(sea_long, aes(x = Conc_nM,y=depth) )+
geom_point() +
facet_grid(Nitrogen~location) + # show an example of change the formula
#facet_grid(~Nitrogen +location)
scale_y_reverse()
ggplot(sea_long, aes(y = Conc_nM,x=as.factor(depth))) +
geom_point() +
facet_grid(Nitrogen~location, scales="free_y")
#Splitting figures using grouping variables with facet_wrap
ggplot(sea_long, aes(y = Conc_nM,x=as.factor(depth))) +
geom_point() +
facet_wrap(Nitrogen~location) +
scale_y_reverse()
ggplot(sea_long, aes(y = Conc_nM,x=as.factor(depth))) +
geom_point() +
facet_wrap(~location + Nitrogen, ncol = 2) +
scale_y_reverse() Exercise 5
#Create a facet_grid plot that does the following:
#1 Has concentration on the y-axis
#2 Is faceted by sample location
#3 Has clearly labeled axes
#4 An improved theme
#5 Different colors for each sample depth
sea_long$Nitrogen<-factor(sea_long$Nitrogen,levels=c("NO3", "NH4"))
# New facet label names for dose variable
loc.labs <- c("Coastal", "Offshore", "Open Ocean", "Gyre")
names(loc.labs) <- c("coastal", "offshore", "openOcean","gyre")
ggplot(sea_long, aes(y = Conc_nM,x=Nitrogen)) +
geom_point(aes(col=as.factor(depth))) +
facet_grid(~location,
labeller = labeller(location = loc.labs)) +
xlab("Nitrogen Species") + ylab("concentration (nM)") +
theme_bw() + labs(color = "Water Depth (m)") +
theme(strip.background =element_rect(fill="white"))+
theme(strip.text = element_text(colour = 'Black',face = "bold"))
Create overlapping data figures with reshaped data
#Compare the distribution of data between groups
ggplot(sea_long, aes(x = Conc_nM, fill = Nitrogen)) +
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Nitrogen Species (nM)") +
scale_fill_discrete(name = "Nitrogen Species",
labels = c("NO3", "NH4")) #be careful here because you have now reordered the levels
#Customize the colors
cols<-c("red","blue")
#cols <- c("#F76D5E", "#72D8FF")
ggplot(sea_long, aes(x = Conc_nM, fill = Nitrogen)) +
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Nitrogen Species (nM)") +
scale_fill_discrete(name = "Nitrogen Species",
labels = c("NO3", "NH4")) +
scale_fill_manual(values=cols)## Scale for 'fill' is already present. Adding another scale for 'fill', which
## will replace the existing scale.
ggplot(sea_long, aes(x = Conc_nM, fill = Nitrogen)) +
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Nitrogen Species (nM)",expand=expansion(mult=c(0.01,.01))) +
scale_y_continuous(expand=expansion(mult=c(0.01,.01)))+
scale_fill_discrete(name = "Nitrogen Species",
labels = c("Nitrate", "Ammonium")) +
scale_fill_manual(values=c("black","lightgray")) +
theme_bw()## Scale for 'fill' is already present. Adding another scale for 'fill', which
## will replace the existing scale.
#Focus on one variable
ggplot(subset(sea_long,Nitrogen == "NH4"), aes(x = Conc_nM, fill = location)) + #could also create a new subsetted dataframe
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Ammonium Concentration (nM)",expand=expansion(mult=c(0.01,.01))) +
scale_y_continuous(expand=expansion(mult=c(0.01,.01)))+
scale_fill_discrete(name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre")) +
theme_bw()
Specify axis breaks
#could specify the breaks we want exactly
ggplot(subset(sea_long,Nitrogen == "NH4"), aes(x = Conc_nM, fill = location)) + #could also create a new subsetted dataframe
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Ammonium Concentration (nM)",expand=expansion(mult=c(0.01,.01)),
breaks=c(0,50,100,150,200,250,300,350,400,450)) +
scale_y_continuous(expand=expansion(mult=c(0.01,.01)))+
scale_fill_discrete(name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre")) +
theme_bw()
#more efficiently, we could use the sequence function, seq()
ggplot(subset(sea_long,Nitrogen == "NH4"), aes(x = Conc_nM, fill = location)) + #could also create a new subsetted dataframe
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Ammonium Concentration (nM)",expand=expansion(mult=c(0.01,.01)),
breaks=seq(0,450,by=50)) + #notice behaviour if we change past the natural x-limits
scale_y_continuous(expand=expansion(mult=c(0.01,.01)))+
scale_fill_discrete(name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre")) +
theme_bw()
ggplot(subset(sea_long,Nitrogen == "NH4"), aes(x = Conc_nM, fill = location)) + #could also create a new subsetted dataframe
geom_density(alpha = 0.75)+
scale_x_continuous(name = "Ammonium Concentration (nM)",expand=expansion(mult=c(0.01,.01)), limits = c(0,500),
breaks=seq(0,500,by=50)) + #notice behaviour if we change past the natural x-limits
scale_y_continuous(expand=expansion(mult=c(0.01,.01)))+
scale_fill_discrete(name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre")) +
theme_bw()
#check the range to make sure we are capturing all the data
range(subset(sea_long,Nitrogen == "NH4")$Conc_nM)## [1] 0.5491473 449.6310512Exercise 6
#Update your sea_long dataset to include the bacteria and viral counts in long format
#Create a facetted plot of the microbial counts by location , colored by depth, and has improved labels and colors
sea_long2<-pivot_longer(sea_long,
cols = c("Bact","Virus"),
names_to="Microbe",
values_to="Counts")
head(sea_long2)## # A tibble: 6 × 9
## Row.names location depth NH4_stat NO3_stat Nitrogen Conc_nM Microbe Counts
## <I<chr>> <fct> <dbl> <chr> <chr> <fct> <dbl> <chr> <int>
## 1 S001 coastal 10 replete depleted NH4 379. Bact 3210498
## 2 S001 coastal 10 replete depleted NH4 379. Virus 35013025
## 3 S001 coastal 10 replete depleted NO3 517. Bact 3210498
## 4 S001 coastal 10 replete depleted NO3 517. Virus 35013025
## 5 S002 coastal 10 replete depleted NH4 312. Bact 3607997
## 6 S002 coastal 10 replete depleted NH4 312. Virus 36342384ggplot(sea_long2, aes(y = Counts,x=Microbe)) +
geom_point(aes(col=as.factor(depth))) +
facet_grid(~location,
labeller = labeller(location = loc.labs)) +
xlab("Microorganism") + ylab("Abundance (particle/ml)") +
scale_color_brewer(palette = "Set2")+
theme_bw() + labs(color = "Water Depth (m)") +
theme(strip.background =element_rect(fill="white"))+
theme(strip.text = element_text(colour = 'Black',face = "bold"))
#Extra if we want to improve the look of the points to not lay on top of each other
ggplot(sea_long2, aes(y = Counts,x=Microbe)) +
geom_jitter(aes(col=as.factor(depth))) +
facet_grid(~location,
labeller = labeller(location = loc.labs)) +
xlab("Microorganism") + ylab("Abundance (particle/ml)") +
scale_color_brewer(palette = "Set2")+
theme_bw() + labs(color = "Water Depth (m)") +
theme(strip.background =element_rect(fill="white"))+
theme(strip.text = element_text(colour = 'Black',face = "bold"))
#Want each microbe on own scale
ggplot(sea_long2, aes(y = Counts,x=location)) +
geom_jitter(aes(col=as.factor(depth))) +
facet_wrap(~Microbe,
labeller = labeller(location = loc.labs),scales = "free_y") +
xlab("Location") + ylab("Abundance (particle/ml)") +
scale_color_brewer(palette = "Set2")+
theme_bw() + labs(color = "Water Depth (m)") +
theme(strip.background =element_rect(fill="white"))+
theme(strip.text = element_text(colour = 'Black',face = "bold"))
Boxplots & violin plots
ggplot(sea_long2, aes(x = location, y = Counts)) +
facet_wrap(~Microbe,scales="free_y")+
geom_boxplot()
ggplot(sea_long2, aes(x = location, y = Counts)) +
facet_wrap(~Microbe,scales="free_y")+
geom_violin()
#can combine the two
ggplot(sea_long2, aes(x = location, y = Counts)) +
facet_wrap(~Microbe,scales="free_y")+
geom_boxplot() +
geom_violin(alpha=0.2)
ggplot(sea_long2, aes(x = location, y = Counts)) +
facet_wrap(~Microbe,scales="free_y")+
geom_violin() +
geom_boxplot(alpha=0.2, width=0.2)
#Exercise 7
#Improve the labels, colors , and theme of the combined box & violin plot
bp<-ggplot(sea_long2, aes(x = location, y = Counts,fill=location)) +
facet_wrap(~Microbe,scales="free_y")+
geom_violin(alpha=0.25) +
geom_boxplot(alpha=0.75, width=0.2)
bp + scale_fill_brewer(palette = "Dark2",name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre")) +
theme_bw() + labs(color = "Location") +
theme(strip.background =element_rect(fill="white"))+
theme(strip.text = element_text(colour = 'Black',face = "bold")) +
scale_x_discrete(name = "Location",
labels = c("Coastal", "Off Shore","Open Ocean","Gyre"))+
scale_y_continuous(name="Abundance (cell/ml)")+
guides(x=guide_axis(angle=45))
#if i waned to change the y-axis numbers https://stackoverflow.com/questions/10762287/how-can-i-format-axis-labels-with-exponents-with-ggplot2-and-scales